By Asa Fitch

If you're already anxious about the AI spending bubble and its global economic risks, maybe don't read the latest annual report from the Bank for International Settlements.

The report, released Monday, lays out in unsettling detail how an AI bust could throw the global financial system into disorder. It also comes from an institution that has a good track record of predicting problems, including the 2008 financial crisis.

This time, the bust could start with overinvestment in AI infrastructure from tech-company spenders. Once it becomes clear that returns aren't measuring up to the scale of those outlays, a pullback in financing and stock prices could ensue.

There's nothing especially new about those concerns. But the BIS connects some other dots, drawing lines between a possible AI bust and the unusually vulnerable state of consumers, governments and the global economy.

U.S. households are significantly more exposed to stocks than in the past couple of decades, for example, both relative to their wealth and their income. And rising U.S. stock prices during the AI boom have given U.S.-listed companies extraordinary weight in global indexes, threatening broad global erosion of wealth in the event of a U.S. tech-focused bust.

Tech companies are only the tip of a supply-chain iceberg, the BIS points out. If they slow the pace of their capital spending, construction contractors with comparatively weak balance sheets could feel the pain quickly, among other financial dominoes ready to fall.

The debt that has come to increasingly fuel AI investments could only compound the potential destruction. And government budgets in advanced economies are stretched, leaving countries less latitude to reinvigorate economic activity through spending. Persistently elevated inflation and geopolitical uncertainty caused by the Iran war make it trickier for central banks to find a good response to any disruption.

These warnings are worth listening to in part because of where they come from. The BIS, a consortium for global central banks based in Switzerland, has raised the alarm early about crises in the past. In March of 2006 for instance, it put out a paper titled "Prime or Not So Prime" that detailed the risks posed by the securitization of subprime loans in the U.S.

That turned out to be prescient. And there's a reasonable chance the BIS is right again about AI, which has been in what looks like a spending and chip-company revenue bubble for some time. Exactly when and how it pops, though, is much harder to guess.

This is an edition of the WSJ AI & Business newsletter, a weekly digest to help you make sense of AI's impact on business with news, insights and data from our global team of technology journalists. If you're not subscribed, sign up here.

South Korean Chip Makers Plan $520 Billion Memory Factory Hub

What AI bubble? So say South Korea's memory chip giants Samsung Electronics and SK Hynix. The companies are planning large new chip-making operations in the country's southwest, substantially expanding their production capacity at a time of unprecedented AI-driven profits. The envisioned plants won't start production for at least several years — at which point the companies hope there is still runaway demand for their products.

The Number

The amount by which Microsoft increased the price of its 1-terabyte Xbox gaming system, responding to rising memory prices being driven by AI demand. Apple also raised prices on iPhones and Macs due to memory costs.

What the Humans Are Saying

AI in Charts

Micron's latest quarterly report confirmed that the memory chip shortage will last well into next year — and likely beyond. The sky-high prices buyers are having to stomach are great for the bottom line of memory chip makers; Micron now makes more operating income in a single quarter than it has ever has reported in full-year revenue up to this point. But those prices and the duration of the shortage have more companies thinking up ways to rely less in the future on the critical component.

AI in the Wild

Axon Enterprise, the maker of Tasers and body cameras, is betting big that AI is the future of policing. Among the tech the company is exploring is software that generates police reports from body-camera audio and a tool that answers policy questions during arrests. Some of Axon's more futuristic ideas have been controversial, though, including a Taser-equipped drone meant to stop school shootings.

Other Highlights From the Week in AI

  • AI server maker Super Micro Computer is in more hot water with Taiwanese authorities over alleged unauthorized diversion of its equipment to China.
  • The U.S. eased some restrictions on Anthropic's most advanced AI models, allowing access to trusted companies and government agencies.
  • Anthropic's competitor OpenAI also said it was limiting access to its most advanced models to a select group of customers approved by the U.S. government.
  • Chinese AI developer DeepSeek is in expansion mode, with plans to double the size of its workforce.

About Us

WSJ AI & Business is a weekly look at AI's transformation of the business world. This newsletter was curated and edited by Dan Gallagher and Asa Fitch. Reach them at dan.gallagher@wsj.com and asa.fitch@wsj.com (if you're reading this in your inbox, you can just hit reply). Got a tip for us? Here's how to submit.